Social Media-Delivered Patient Education in Enhancing Type 2 Diabetics Self-Management and Attitudes During the COVID-19 Pandemic

The study was conducted in the Endocrinology and Metabolism Clinic in Taipei Medical University Wang-Fang Hospital. Patients were referred from physicians if they met the inclusion criteria and consent to join. Written consent and questionnaires were then provided to the participants after researcher had clearly explained the study purpose.

Based on the seven key points in managing diabetes developed by the association of diabetes care and education specialist (ADCES), including healthy coping, healthy eating, being active, monitoring, taking medication, problem solving and reducing risks, the research team had developed 51 diabetes related health educational videos (available at website of School of Pharmacy, Taipei Medical University http://pharmschool.tmu.edu.tw/activity/index.php?type=20). Each video lasted for about 2-3 minutes. According to the content of the videos, it was categorized into 5 domains, including understanding diabetes, daily care, nutrition care, diabetes drug and diabetes knowledge related quizzes.

Patients in intervention group could attend the video through the line platform. Researcher would also send 2-3 videos per week with care massage every 2 weeks to the patients. The general schedule of video was shown in table x. According to various condition, researcher developed specific video schedule for each patients. For example, patient who was prescribed with A drug, we would only send the video regarding A drug to this patient but not all of the video regarding drugs. Schedule for the part of understanding diabetes, daily care, nutrition care and quizzes was the same among all participants. Due to the quizzes that were put in week 3, 6, 9 and 12, videos related to general knowledge about diabetes would be send to patients first to make sure all the quizzes related content had already taught to patients before the quizzes. Moreover, patients could communicate with the research team and ask questions through the platform, researcher would provide answers verified by pharmacists or physicians in one or two days after.

The sample size of this study was estimated using G-power (version 3.1). Assuming a power of 80% with two-sided alpha level for detecting the difference between the intervention group and the control group, at least 64 patients were needed for each group. Considering 20% dropout rate, the study was designed to have at least 80 patients for each group, 160 patients in total.

Patients were randomized in a 1:1 ratio according to the random allocation sequence generated prior to the study. Patients with odd number were allocated into control group while even numbers were allocated into intervention group. The trial was non-blinded because of its feasibility.

Researcher would not provide any inductive explanation to the patient during filling the questionnaire. If patient had questions about the items, a detailed explanation and health education were provided only after finishing the post-test. The returned questionnaires were collected by the researcher and kept in the Department of Clinical Pharmacy, School of Pharmacy, Taipei Medical University for three years.

All the data were analyzed using SPSS (SPSS Inc. Released 2009. Predictive Analytics Suite Workstation (PASW) Statistics for Windows, Version 18.0. Chicago: SPSS Inc.). A two-tailed p value of 0.05 was considered to be statistically significant in the analysis. For personal characteristics, descriptive analysis was performed. According to the characteristic of data, Chi-square test and t-test was used for detecting the differences between groups. a Wilcoxon signed rank test or paired t-tests was performed for the differences in A1C and the scores between pre and post-test about knowledge, self-care activity and attitude towards diabetes. A Mann-Whitney test or unpaired t-tests was performed for the mean differences between control and intervention group. Logistics regression was performed for the correlation between variables.

Patients who had finished both pre- and post-assessment would be included for analysis, those without post-test would be excluded. Missing values were processed by multiple imputation.